Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
SECTION 1
APPROACHES TO EDUCATIONAL RESEARCH
Educational research can take a wide variety of forms and serve many different
purposes. Sometimes it is located within social science disciplines, especially
psychology and sociology. Other work is of a more policy-oriented nature, and
some is quite closely linked to educational practice. There is also diversity at the
level of methods used. These include: laboratory and classroom experiments; large-
scale surveys of the behaviour, attitudes, aptitudes, etc., of teachers, children, heads,
governors and others; analysis of published and unpublished texts,-both qualitative
and quantitative; and small-scale investigations of particular institutions or locales.
A common way of conceptualizing this diversity in method and data is the
distinction between quantitative and qualitative approaches, between research
which relies primarily on numerical data and that which uses mainly verbal data.
This is a distinction which will be used in this module; though as we shall see it is
by no means
Psychological approaches
As de Landsheere makes clear, the new educational research had its origins in late
nineteenth century psychology, which was itself only just emerging as a distinct
discipline at the time. This initial connection with psychology affected both the
topics that were investigated and the methods which were used. Above all, it led to
a commitment on the part of educational researchers not just to a scientific
approach to their work but also to a particular interpretation of the methodological
requirements of science; one which has come to be referred to broadly, and
somewhat as 'positivist'. This placed great emphasis on the need for
quantitative measurement of the characteristics of learners and teachers, and of their
behaviour. Experimental method was regarded as the ideal model of a scientific
approach, even though by no means all early educational research was
experimental in character. The aim of the new educational research was to lay a
Section 1 Approaches to educational research 11
theoretical basis for understanding the processes of teaching and learning, one
which would revolutionize education by putting it on a scientific footing.
A particularly important aspect of this early history of educational research was the
construction of mental tests of various kinds - of intelligence, academic
achievement, personality, and attitude — which were applied to pupils. Alongside
these were developed various rating scales and observational schedules for
measuring aspects of teachers' behaviour (Dunkin and Biddle, 1974). Such tests and
behavioural measurements were believed to offer teachers, educational
administrators and others valuable information of an objective kind which would .
facilitate both effective educational planning and the monitoring of educational
processes to assess their success.
This sociological research was not experimental in character: it involved the analysis
of official statistics and of data from questionnaire surveys. However, it employed
similar measurement techniques (for instance, of pupils' ability and social class) to
those used in psychological research. It also used statistical analysis designed to
simulate the manipulation of variables involved in experimental research, and
thereby to identify causal relationships. Quantitative research relying on these
techniques continues today though now the focus is more likely to be on
differences in educational outcomes between pupils from different ethnic groups
or on 'school effectiveness' (see, for instance, Drew and Gray, 1990 and 1991; Gray
et al., 1990). As an illustration, we will look briefly at the latter.
School effectiveness
The measurement of 'school effects' is an area of educational research which has
come to have considerable significance in recent years. One of the original stimuli
for this research was the report in the United States by Coleman (1966). This was
See Floud et al., 1956, and Douglas, 1964. A recent study in the same tradition, which
brings the story a little more up to date, is Halsey et al., 1980. For similar studies relating to
Scotland and Northern Ireland, see Gray et al., 1983 and Cormack and Osborne, 1983,
respectively.
12 E835 Educational Research in Action Part 1
Educational evaluation
Another area where a quantitative approach became very influential was in the field
of educational evaluation. In Britain in the 1960s many large-scale projects for
curriculum development were sponsored by the Schools Council and by private
funding agencies, such as the Nuffield Foundation. Sometimes, these projects were
subject to evaluation as part of the process of implementation. Initially, this often
took the form of a translation of the objectives of the project into quantitative terms,
and investigation relying on measurement of pupils' achievements and attitudes to
determine whether those objectives had been met. This quantitative approach to
evaluation has continued to be influential; though, as we shall see, qualitative forms
of evaluation have become more prominent in recent years.
In our brief and rather selective account of the history of educational research we
have concentrated on what, until fairly recently, was the most influential element of
it: quantitative approaches modelled to one degree or another on what was taken to
be the method of natural science. We turn next to qualitative research, which has
come to have considerable influence in the past few years.
Measurement problems
Quantitative researchers themselves are not unaware of these problems, but they
view them rather differently from their qualitative critics. We can get a sense of this
by looking at an influential commentary on some experimental research by the
famous Swiss psychologist Jean Piaget. Over a lifetime of work, which spanned
much of the twentieth century, Piaget developed an influential account of child
development that portrayed the child as evolving through various stages, primarily
as a result of practical involvement with the physical and social world, each stage
providing for progressively more complex capabilities. His work has been criticized
by some for being insufficiently rigorous from an experimental point of view. In part
this stems from a difference between Piaget and his critics about the requirements of
scientific research. But it also reflects an increasing awareness by psychologists of
the extent to which their experimental findings are open to alternative
interpretations, and their attempts to design more sophisticated experiments to
allow for these.
Piaget's interpretation of his research was, of course, that most of the children were
unable to perform the logical task involved in answering his questions, this being
because their cognitive development had not reached the necessary stage.
Donaldson questions this. She mentions the possibility that the children were
simply unwilling to play the experimenter's game, but she concentrates particularly
on the suggestion that the children misunderstood what the experimenter was
asking. This points to the fact, obvious enough but important in its implications, that
experiments are social situations in which interpersonal interactions take place. The
implication of this is that Piaget's work, and attempts to replicate it, not only involve
14 E835 Educational Research in Action Part 1
measurement of children's capacities for logical thinking but also of the extent to
which they have understood what was required, their willingness to comply with
these requirements, and the experimenters' success in communicating what is
wanted and motivating the children to answer appropriately.
The response of Donaldson and her co-researchers to these problems was to devise
ingenious new experiments which sought to test the competing interpretations to
which Piaget's original work was subject. Others have taken a more radical line,
however, treating this type of problem with experimental research as one which
cannot be overcome by an improvement in technique. Instead, it has been taken to
indicate that there is something fundamentally wrong with experimental research,
or at least that it suffers from severe limitations.2
Parallel arguments emerged in relation to other forms of quantitative research. Thus,
we find Mehan developing similar criticisms of psychological and educational tests
to those which Donaldson made of Piaget's work. He points out how test questions
may be interpreted in ways that are quite different from those intended by the
researcher:
A question from [a] language development test instructs the child to
choose the 'animal that can fly' from a bird, an elephant, and a dog. The
correct answer (obviously) is the bird. Many first grade children, though,
chose the elephant along with the bird as a response to that question.
When I later asked them why they chose that answer they replied:
That's Dumbo'. Dumbo (of course) is Walt Disney's flying elephant, well
known to children who watch television and read children's books as an
animal that flies.
(Mehan, 1973, p. 249)
Here Mehan is pointing out that interpretation of the results of tests depends on the
assumption that there is a correspondence between the interpretative frame
employed by the test constructor and that adopted by those who take the test. Any
disparity between these frameworks may mean that the test is not measuring what it
is intended to
Test constructors recognize this problem and engage in a great deal of pilot research
to eliminate potential ambiguities and misunderstandings. In other words, they see
the problem as a technical one which can be minimized by improved test
construction. However, Mehan believes that the problem is endemic to tests, and
that it can only be dealt with by a change in approach. He summarizes his argument
as follows:
This examination of testing interactions shows test assumptions are not
met in practice [...] Test materials do not always have the same meaning
for tester and child. The child's performance is not just the result of his
ability and the stimulus presented but is also influenced by contextually
provided information [...] In short, test taking and test scoring are
interpretive interactional processes which should be approached and
studied as such.
(Mehan, 1973, pp. 255-6)
In other words, Mehan believes that what is required is qualitative research into the
interpretive and interactional processes involved in learning, not reliance on
quantitative measurements whose validity is in doubt.
Much the same criticism of quantitative research was developed in other areas too,
for example in relation to the official statistics and questionnaires on which much
social and educational research relies (Kitsuse and Cicourel, 1963; Phillips, 1976).
Thus, it has been pointed out that survey data do not simply represent facts about
the world, but are the product of complex patterns of social interaction between
interviewers and interviewees.
However, it was not just the failings of psychological, social and educational
measurement which were criticized. Questions were also raised about the
assumption, built into the logic of quantitative research, that causes can be
identified by physical and/or statistical manipulation of variables. Critics suggested
that this fails to take account of the very nature of human social life, assuming it to
consist of mechanical cause-and-effect relationships; whereas, in fact, it involves
complex processes of interpretation and negotiation that do not have determinate
outcomes. From this point of view, it is not clear that we can understand why
people do what they do in terms of the simple sorts of causal relationships on
which quantitative research focuses. Social life, it is suggested, is much more
contextually variable and complex (for these arguments, see, Blumer, 1969; Matza,
1969).
These criticisms of quantitative research are not simply about method, about the
relative effectiveness of different research strategies. They also involve more deep-
seated disagreements about the nature of human behaviour and how it can be
understood. Very often this has been formulated as a conflict between the positivist
assumptions of quantitative research and the very different assumptions, sometimes
referred to as 'naturalistic', 'interpretive' or 'phenomenological', of the newer
qualitative approaches. This alternative philosophy of research stresses the way that
people's perspectives on the world shape their actions, and the diversity of such
perspectives. It has led to an emphasis on the importance of researchers
understanding those perspectives, this often being seen to require an exploratory
approach in which the researcher suspends his or her own assumptions in order to
learn to see the world from others' points of view. Equally important, anti-positivist
views treat action as not being determined by prior perspectives but rather as
constructed over time in ways which are sensitive to their contexts. This results in
an emphasis on the detailed investigation of actual social processes as these occur
in natural situations, instead of (or as a complement to) the use of other types of
data.
Besides defending their work against criticism in this way, quantitative researchers
are also able to point out that qualitative inquiry is itself by no means free of
problems. It often deploys verbal quantifications, for example by means of such
words as 'often', 'usually', 'frequently', 'generally', 'typically', etc., and yet the use of
these is not usually based on any rigorous counting procedure; indeed, there may
not even be any explicit indication of what were and were not taken to be instances
of the phenomena concerned. This means that the generalizations produced are
subject to the sorts of errors that have been documented in informal observation,
such as being unduly influenced by novel or extreme occurrences (Sadler, 1981).
Similarly, qualitative researchers often make causal claims, but their procedures of
inquiry rarely involve the sort of control of variables which experimenters or survey
researchers use; and it is not clear what other means are effective in ruling out
alternative explanations.
There is much truth in both the criticisms and the counter-criticisms. However, the
contrast assumed by this debate between two paradigms, between two opposed
methods and philosophies of social and educational research, is rather misleading.
Much educational inquiry uses both quantitative and qualitative methods, and there
are good reasons for believing that these can complement one another in important
ways (see Bryman, 1988 and 1992). Furthermore, at the philosophical level, while
positivism has certainly been very influential in the twentieth century, not least in
promoting the view that natural science is the model for all knowledge and inquiry,
there has been considerable disagreement amongst those influenced by this
philosophy about the nature of scientific method and about how it should be
applied to the study of the social world. Nor is there just one alternative to
positivism, so that the influence of many different forms of anti-positivism is to be
found within qualitative research. Equally important, divergence in philosophical
orientation is not all that divides qualitative researchers; there are also differences
stemming from disciplinary and theoretical commitments, and from different ideas
about the proper relationship between research and practice.
In the late 1960s and early 1970s other qualitative researchers within sociology
broke more sharply with the earlier tradition of quantitative educational sociology.
These 'new sociologists of education' (Young, 1971; Gorbutt, 1972) argued that this
research did not ask deep enough questions about the phenomena it investigated,
that it took too much for granted. For instance, it assumed that the education which
schools dispensed was of positive value and therefore it did not give enough
attention to the nature of school knowledge and pupils' learning, concentrating
exclusively on the distribution of educational opportunities. These 'new
sociologists' sought to place the question of who defines what constitutes education
on the research agenda. They suggested that the nature of teaching and learning
processes in schools reflects the cultural, and (ultimately) the political and
economic, dominance of some groups in society over others. This change of
theoretical orientation in the sociology of education had methodological
implications: it was widely believed that only qualitative research could provide an
understanding of the cultural and political processes involved in schooling.
Both the example of Hargreaves and Lacey and the writings of the 'new sociologists'
encouraged the growth of ethnographic and other forms of qualitative research in
the 1970s and 1980s. At the same time there was a shift in interest away from social-
A girls' grammar school was also studied at the same time by Lambart, but the results were
not published until later: see Lambart, 1976 and 1982.
18 E835 Educational Research in Action Part 1
class inequalities to those relating to gender and 'race'. This was stimulated, in large
part, by the growing influence of feminism and of multiculturalism/anti-racism.
Many of the same arguments about the role of schools in generating inequalities
were developed here, and much emphasis was placed on qualitative investigation
of school processes. One consequence of this has been an increased amount of
research of a qualitative kind on women's and girls' experiences of the educational
system (see, for example, Stanworth,1981; Griffin, 1985; Arnot and Weiner, 1985;
Weiner and Arnot, 1987). Similarly, there has been a growing body of research
looking at the experience of ethnic-minority children in schools (see, for instance,
Eggleston et al, 1986; Mac an Ghaill, 1988; Foster, 1990; Gillbom, 1990).5
Foster et al., 1995, provide a critical assessment of research on school processes and
educational inequality.
For an illuminating history of the field of curriculum evaluation, see Hamilton, 1976.
Section 1 Approaches to educational research 19
7
For a now classic example of this work in the history and philosophy of science, see Kuhn,
1970.
8
For an account of the development of these arguments, see Hammersley, 1995, Chapter 1.
20 E835 Educational Research in Action Part 1
Can the quest for objectivity distract us from the pursuit of truth?
9
For an account, see Best and Kellner, 1991. For examples of the application of these ideas
within educational research, see Lather, 1991; Ball, 1990a; and Usher and Edwards, 1994.
Section 1 Approaches to educational research 21
thereby gives direct access to reality. It is important to notice, however, that Phillips
also rejects this position.
By contrast, Phillips takes as his target relativism. This relativism, however, is not
synonymous with the relativism that Eisner admits to and, for this reason, we shall
call it 'extreme relativism'. Extreme relativism is the view that all 'knowledge' is a
construction based on a particular framework of presuppositions, that these
presuppositions can never be fully assessed because all assessments themselves rely
on presuppositions, and that all empirical claims must therefore be treated as
equally valid in their own terms. From this point of view we cannot talk of validity
in terms of correspondence to a reality that stands outside of any framework of
assumptions, nor is there a procedure which provides access to any such reality.
We are not faced, then, with a conflict between two positions each represented by
our authors, but rather with two authors attacking opposite polar positions that
neither of them seems to occupy. Thus, Eisner suggests that quantitative educational
research is founded on naive realism. Yet the philosophical ideas associated with
quantitative research have been quite diverse and have included rejection of naive
realism in favour of approaches which seek to avoid all reference to any reality
beyond our experience. Indeed, what Eisner refers to as procedural objectivity has
been regarded by some quantitative researchers as the only form of objectivity there
is, agreeing in this respect with him that this is 'all we can ever have' and that we
must 'recognize it for what it is'. (This is a point that Phillips makes on page 66.)
Similarly, Phillips treats Eisner as effectively claiming that any view is as good as any
other, that this is what the abandonment of objectivity means. Yet Eisner clearly
does not see his position in these terms. Towards the end of his article he quotes the
philosopher of science Stephen Toulmin to the effect that even in the absence of
knowledge that is certain we can still make reasonable assessments of competing
claims. It must be said, though, that Eisner does not spell out how this is to be done
and, in particular, how judgements of validity are to be justified: nor does he
address the issue which Phillips raises about whether it is possible to offer rational
justification for the selection of frameworks.10
These two articles indicate the sorts of philosophical issues which are at the heart of
much discussion about validity among qualitative educational researchers today.
Our analysis of them shows that the differences of view to be found are often more
complex and subtle than they might at first appear.
1.3 CONCLUSION
In this section we have looked at the history of educational research and at some of
the debates about how it can and should be pursued. We examined the dominance
of quantitative approaches and the reaction against them, in the context of debates
about positivism and anti-positivism. And we noted how criticism of quantitative
method was accompanied by a great increase in the amount of qualitative research,
initially in the sociology of education and in evaluation studies, but later across all
fields of educational investigation. At the same time, we saw how the continuing
commitment of much qualitative research to key elements of the scientific method
has recently come under challenge; in part through the influence of poststructuralist
and postmodernist ideas.
We have used the distinction between quantitative and qualitative approaches to
educational research as a way of organizing our discussion up to this point in the
course, and we shall go on using it. However, as we mentioned earlier, there are
dangers with this distinction. It may lead us to believe that there are just two
alternatives in doing research, so that one must choose either to do a quantitative or
a qualitative study. Yet there are different kinds of quantitative research, and there is
10
More extensive presentations of these authors' positions can be found in Eisner, 1991, and
Phillips, 1992.
Section 2 Research, educational practice and politics 23
In the next section we turn to an issue that confronts both quantitative and
qualitative researchers: the question of what the relationship is and should be
between their work and educational practice; and of the extent to which that
relationship is 'political'.